About

Mini AI Lab leverages modern web technologies such as WebGPU, WASM, and WebCodecs to run AI models and multimedia tasks entirely inside your browser without uploading data to servers.

Core Values

  • Privacy First: Photos, audio, and text are processed strictly on your device and are never sent to external servers.
  • Zero Installation: Run directly in modern web browsers without complex installation.
  • Open Standards & Open Source: Built on robust technologies including Transformers.js, ONNX Runtime Web, and WebCodecs.

Available Services

We provide a collection of on-device tools including image resizing, background removal, image captioning, voice-to-text (STT), and vertical video (Shorts) maker.

  • Voice to Text — Upload an audio file and get what was said written out as text.
  • Background Removal — Upload a photo and download it with the background cleanly removed.
  • Describe a Photo — Upload a photo and get a sentence describing what is in it.
  • Photos → Shorts — Pick photos to make a 9:16 vertical Shorts video. Background music is optional.
  • Resize Image — Fit a photo to the Shorts ratio (9:16) or another ratio and download it. Empty areas are filled in automatically.
  • Noise Remover — Reduce the noise in a recording so voices or nature sounds come through clearly. Your file never leaves your device.